Different Year, Different Fish: Tracking Change in Amazon Floodplain Streams

2025 FIELD REPORT: STREAM SURVEY (PDF)

Date: July 14-19, 2024 and July 11-20, 2025
Location: Explorama’s Napo-Sucusari Biological Reserve, Loreto Peru
Study Site: The Amazon Conservatory for Tropical Studies (ACTS). Forest floor.
Lead Researcher: Dr. Marie Trone*
Team Members: (2024): Dina DiSantis, Kathy Hockman, Percy Reyina Inuma, Lucas Kahler, Philip Kahler, Ben Olsen, Kathy Richardson, Cesar Sevillano, Jordan Wolf. (2025): Brice Britttsan Amanda Dworak-Rowland, Kirsten Franklin, Carissa N. Ganong, Erica Groendal, Kathy Hockman, Michael Holmes, Percy Reyina Inuma, Morgan Johnson, Janet Ort, Liana Pellegrino, Kathy Richardson, Callie Rundhammer, Cesar Sevillano, Summer Schultz, Karin Pinchi Villanueva, Victoria Vicente, David Watson. (* = author of this report)

Abstract
Fishes inhabiting the Peruvian Amazon floodplain experience extreme environmental variability driven by annual flood pulses that alter habitat availability and water chemistry. To assess how these dynamics influence local assemblages, fishes and associated macrofauna were sampled at six small streams near the ACTS Field Station during July 2025 and compared with pilot sampling from July 2024. Family-level richness ranged from 1 to 7 families per site (mean = 3), and abundance ranged from 1 to 19 individuals (mean = 6.7), with eight fish families, two crustaceans, and one mollusk recorded. Dissolved oxygen concentrations were uniformly low to moderate (0.6–5.63 mg/L), and neither dissolved oxygen nor stream width or depth was significantly correlated with species richness or abundance. In contrast, pronounced interannual differences in assemblage composition were observed, including the presence of splashing tetras, hatchetfishes, and tadpoles in 2024 but not in 2025, and the appearance of electrogenic gymnotid fishes and vocal doradid catfishes in 2025. These results indicate that snapshot abiotic conditions weakly predict assemblage structure, whereas flood-pulse–driven temporal variability plays a dominant role in shaping Amazonian floodplain fish communities.

Lay/Marketing Summary
Fish communities in the Peruvian Amazon floodplain experience strong seasonal and interannual variability driven by annual flood pulses. To examine how these dynamics influence local assemblages, Morpho researchers sampled fishes and associated macrofauna in six small forest streams near the ACTS Field Station in July 2025 and compared results with pilot sampling from July 2024.

Across sites, fish family richness and abundance were low to moderate and showed no significant relationship with dissolved oxygen, stream width, or depth. In contrast, assemblage composition differed markedly between years. Several taxa observed in 2024—including splashing tetras, hatchetfishes, and tadpoles—were absent in 2025, while electrogenic knifefishes and vocal doradid catfishes appeared in 2025 but not the previous year.

These findings suggest that snapshot abiotic measurements weakly predict community structure in Amazon floodplain streams. Instead, flood-pulse–driven temporal variability plays a dominant role in shaping fish assemblages, underscoring the importance of multi-year sampling to understand biodiversity patterns in highly dynamic tropical river systems.

Introduction
Fishes living in the Peruvian Amazon floodplain face a suite of intense and recurring challenges driven by the region’s annual flood pulse. Dramatic water-level fluctuations of several meters alternately expand and contract available habitat, transporting, concentrating, and isolating fishes each year (Sousa et al., 2022). During low-water periods, fishes contend with extremely low dissolved oxygen, high turbidity, elevated temperatures, and acidic water (low pH), conditions that limit aerobic metabolism, disrupt ion regulation, and increase physiological stress. As waters recede, habitats become fragmented, intensifying competition, predation, and disease transmission while reducing food availability (Anjos et al., 2008; Borghezan et al., 2021; Luis Val & Wood, 2022). Seasonal shifts in water chemistry, unstable substrates, and the need to synchronize reproduction with flood timing further challenge survival (Anjos et al., 2008; Borghezan et al., 2021; Luis Val & Wood, 2022). Together, these pressures have shaped a fauna characterized by remarkable physiological, behavioral, and morphological adaptations to life in one of the most dynamic freshwater environments on Earth (Anjos et al., 2008; Luis Val & Wood, 2022).

Many fish species are hypoxia-tolerant, lowering their metabolic rate to reduce oxygen demand, while others supplement gill respiration by gulping air or using modified organs such as lungs, swim bladders, or vascularized mouth and gut tissues (Braz‐Mota & Almeida‐Val, 2021). These strategies also allow fishes to use low-oxygen habitats as refuges from predators. Some fishes generate and sense weak electric fields to navigate, communicate, and detect nearby organisms in dark, turbid waters where vision is limited (Crampton, 2019). Many species also communicate using sound; for example, catfishes (Order: Siluriformes) produce sounds through stridulation and by vibrating the gas bladder with specialized muscles to signal during territorial disputes and reproduction (Borghezan et al., 2021). Anti-predator adaptations include both behavioral and morphological strategies. For example, splashing tetras (Family: Characidae) deposit eggs on terrestrial vegetation above the waterline (Farias et al., 2025) potentially reducing predation. Hatchetfishes (Family: Gasteropelecidae) possess a laterally compressed, keel-shaped body that enhances rapid surface bursts and aerial escape (Weist, 1995). Together, these pressures have driven extensive specialization, contributing to the Amazon Basin’s extraordinary fish diversity of over 3,000 freshwater species—nearly 20% of the world’s total.

The ACTS Field Station typically reaches peak flood stage in May–June, with water levels receding by July (Database for Hydrological Time Series of Inland Waters [DAHITI], n.d.). The surrounding area contains multiple small streams ranging in depth from a few centimeters to a maximum of 1–2 meters. Pilot sampling conducted in July 2024 at four sites indicated no significant differences in species composition among traps baited with a control (no bait), fish, or bananas and rice. Based on these findings, the present study resampled two of the original sites to evaluate interannual changes in species assemblages and expanded sampling to include four additional sites.

Materials and Methods
Fish were sampled from six bridges that crossed streams located along trails surrounding the ACTS Field Station (Figs. 4–5) daily from 14–17 July 2025. At each site, traps were suspended with a portion extending above the water surface to permit access to atmospheric air for hypoxia-adapted species. Traps were baited with a mixture of dorado fish, beef, and bananas

Two traps were deployed at each bridge between 1500 and 1700 h, allowed to trap fish overnight, and retrieved the following morning between 1000 and 1200 h. Captured fishes were transported to the field laboratory, where they were photographed and measured, then returned alive to their original collection sites.

Acoustic signals produced by two catfish species were documented opportunistically using a Cetacean Research SQ26 hydrophone, sampled at 256 kHz using a Qualilife Highblue v3.0 acoustic acquisition system (Glotin, Toulon University), paired with synchronous video recording (Fig. 6). These recordings were analyzed using Raven Pro 1.4.

At each sampling site, stream width and depth were measured. Dissolved oxygen and pH were recorded using a PASCO Optical Dissolved Oxygen Sensor and a PASCO pH Sensor (PS-2102). Panoramic video was collected to document surrounding vegetation, and GPS coordinates were recorded at all sites using a Garmin inReach.

Fish were identified to the family level; crustaceans and mollusks were also documented when present.

Results
Across 6 sampling sites, family-level richness ranged from 1 to 7 families per site (mean = 3), and total abundance ranged from 1 to 19 individuals (mean = 6.7). In total, 8 fish families were recorded, along with 2 crustacean species and 1 mollusk species (Fig. 7). Quebrada Tucanes (Baños) exhibited the highest richness and abundance (7 families; 19 individuals; Fig. 8), whereas Quebrada Chinganero and Quebrada Titi were each represented by a single family and individual (Fig. 9). Quebrada Tucanes (Baños) measured 305 cm in width and 32 cm in depth, while Quebrada Chinganero measured 850 cm in width and 149 cm in depth (Table 1).

At Quebrada Puentes, traps were initially deployed during abiotic data collection and subsequently repositioned along the bridge to ensure that a portion of each trap remained above the water surface, allowing air-breathing and hypoxia-tolerant organisms access to the surface. Approximately 17 fish entered the traps during the initial 15-minute deployment; however, family composition and body size were not documented at that time so trapping could proceed overnight under standardized conditions. Upon retrieval the following day, only 5 individuals from the family Characidae were captured, along with 1 crab and 1 snail.

Dissolved oxygen concentration was not significantly correlated with fish abundance (Pearson r = −0.09, p = 0.87; Spearman ρ = 0.12, p = 0.83; Fig. 10). Dissolved oxygen measured 0.6 mg/L at Quebrada Chinganero, where a single fish was captured, and 5.63 mg/L at Quebrada Titi, where only 1 crab was collected. Quebrada Tucanes (Baños), which had the highest richness and abundance, had a dissolved oxygen concentration of 3.17 mg/L (Table 1; Fig. 10).

All sites were near neutral pH (6.99–7.01) and were characterized by standing water with no measurable flow (Table 1).

Acoustic signals were recorded from 2 catfish individuals belonging to the family Doradidae. One individual produced sounds via stridulation, whereas the other produced low-frequency swim bladder–mediated drumming sounds (Fig. 13).

Discussion
Quebrada Tucanes (Baños) exhibited the highest species richness and abundance, with 7 of the 8 families captured at this site, despite a dissolved oxygen concentration of 3.17 mg/L. Quebrada Chinganero and Quebrada Titi were each represented by a single family and individual. These 2 sites were characterized by the lowest and highest dissolved oxygen concentrations measured (0.6 mg/L and 5.63 mg/L, respectively; Fig. 7; Table 1).

Healthy freshwater fish communities typically require dissolved oxygen concentrations of approximately ≥ 4–6 mg/L for routine metabolic function, whereas concentrations below ~2–3 mg/L are broadly considered hypoxic and limiting to aerobic aquatic life (Diaz & Rosenberg, 2008). Most sampled sites fell within hypoxic or near-hypoxic conditions, and dissolved oxygen concentration alone did not appear to influence species richness or abundance in this system.

Similarly, stream width and depth did not appear to influence species richness or abundance. The site with the highest richness and abundance was characterized by intermediate width and depth, whereas the sites with the lowest abundance (each represented by a single individual) occurred at both extremes of habitat size—one being the widest and deepest site and the other among the narrowest and shallowest (Figs. 11, 12). However, because only a single sample was collected at each of the 6 sites, statistical power for detecting relationships between habitat dimensions and assemblage patterns was limited.

Quebrada Tucanes (Baños) supported the greatest diversity, including electrogenic and electroreceptive fishes (Gymnotidae) (Fig. 3) and predatory fishes from the family Erythrinidae (Fig. 14). Despite this high biodiversity, the same site contained only 5 individuals representing 3 families during sampling in the previous year, with only 1 family occurring in both years (Figs. 7, 15). Quebrada Puentes, which was also sampled in both July 2024 and July 2025, showed similarly pronounced interannual differences in faunal assemblages. In July 2024, 21 individuals representing 4 families were captured, including 2 crabs and 3 tadpoles, whereas sampling in July 2025 yielded 7 individuals representing 3 families, including 1 crab and 1 snail (Figs. 7, 15).

Notably, 5 splashing tetras (family Lebiasinidae) (Fig. 1) were captured at Quebrada Puentes and 1 at Quebrada Tucanes (Baños) in 2024, yet none were recorded at any of the 6 sampling sites in 2025 (Figs. 7, 15). These fishes are notable for depositing eggs on terrestrial vegetation and possessing a heterocercal caudal fin (Fig. 1), which facilitates repeated jumping and splashing behavior used by males to keep eggs moist until hatching occurs (Farias et al., 2025).

Fish diversity was further illustrated by the capture of 6 hatchetfish in July 2024 (Fig. 2), with none recorded in July 2025. In contrast, species captured in July 2025 but not in 2024 included 2 electrogenic and electroreceptive gymnotid fishes and 2 vocal catfishes (family Doradidae), each producing distinct acoustic signals via different sound-production mechanisms—stridulation and swim bladder–mediated drumming (Figs. 6, 13) (Borghezan et al., 2021).

Together, the pronounced interannual shifts in species richness and assemblage composition observed at Quebrada Tucanes (Baños) and Quebrada Puentes support the role of annual flood-pulse dynamics in restructuring floodplain fish and macrofaunal communities, resulting in substantial year-to-year turnover in species presence and abundance (Anjos et al., 2008; Borghezan et al., 2021; Luis Val & Wood, 2022. Sousa et al., 2022).

Conclusion
Overall, patterns of species richness and assemblage composition across sites were not strongly predicted by dissolved oxygen concentration or physical habitat dimensions, but instead reflected pronounced interannual variability. The substantial year-to-year turnover observed at multiple sites is consistent with flood-pulse–driven redistribution of fishes and macrofauna in Amazonian floodplain systems, highlighting the dynamic nature of these habitats and the importance of future temporal sampling for understanding community structure.

References
Anjos, M. B. D., De Oliveira, R. R., & Zuanon, J. (2008). Hypoxic environments as refuge against predatory fish in the Amazonian floodplains. Brazilian Journal of Biology, 68, 45-50.

Borghezan, E. D. A., Pires, T. H. D. S., Ikeda, T., Zuanon, J., & Kohshima, S. (2021). A review on fish sensory systems and Amazon water types with implications to biodiversity. Frontiers in Ecology and Evolution, 8, 589760.

Braz‐Mota, S., & Almeida‐Val, V. M. (2021). Ecological adaptations of Amazonian fishes acquired during evolution under environmental variations in dissolved oxygen: A review of responses to hypoxia in fishes, featuring the hypoxia‐tolerant Astronotus spp. Journal of Experimental Zoology Part A: Ecological and Integrative Physiology, 335(9-10), 771-786.

Crampton, W. G. (2019). Electroreception, electrogenesis and electric signal evolution. Journal of Fish Biology, 95(1), 92-134.

Database for Hydrological Time Series of Inland Waters (DAHITI). (n.d.). Water level altimetry – Mazan River (ID 32287). DGFI-TUM. Retrieved December 26, 2025, from https://dahiti.dgfi.tum.de/en/32287/water-level-altimetry/

Diaz, R. J., & Rosenberg, R. (2008). Spreading dead zones and consequences for marine ecosystems. science, 321(5891), 926-929.

Farias, R. R., López-Rodríguez, N. C., Gonçalves, L. A., Tavares, C. D. N. D. S., Rocha, R. M., Freitas, T. M. D. S., & Montag, L. F. D. A. (2025). Reproductive biology of Copella arnoldi (Characiformes: Lebiasinidae), a terrestrial-spawning fish from the Amazon. Neotropical Ichthyology, 23(01), e240054.

Glotin, H. University of Toulon. (Need to get a better citation here).

Luis Val, A., & Wood, C. M. (2022). Global change and physiological challenges for fish of the Amazon today and in the near future. Journal of Experimental Biology, 225(10), jeb216440.

Sousa, R. G. C., Oliveira, N. S., & da Rosa, F. R. (2022). The flood pulse regulates the longitudinal distribution of fish assemblages in the amazonian floodplain lakes. Boletim do Instituto de Pesca, 48.

Touchon, J. C., & Worley, J. L. (2015). Oviposition site choice under conflicting risks demonstrates that aquatic predators drive terrestrial egg-laying. Proceedings of the Royal Society B: Biological Sciences, 282(1808), 20150376.

Wiest, F. C. (1995). The specialized locomotory apparatus of the freshwater hatchetfish family Gasteropelecidae. Journal of Zoology, 236(4), 571-592.
References

For additional information or questions about this field report, please contact Dr. Lindsey Swierk (lindseyns@gmail.com), Director of Scientific Research for The Morpho Institute and Associate Director of Research for the Amazon Conservatory for Tropical Studies. 

Learn more about the Amazon Research Initiative for Educators (ARIE)

Download ARIE Syllabus